Journal Pre-proofs

Flexible CIGS, CdTe and a-Si:H based thin film solar cells: A review

Jeyakumar Ramanujam, Douglas M Bishop, Teodor K Todorov, Oki

Gunawan, Jatin Rath, Reza Nekovei, Elisa Artegiani, Alessandro Romeo

PII: S0079-6425(19)30101-X
DOI: https://doi.org/10.1016/j.pmatsci.2019.100619
Reference: JPMS 100619

To appear in: Progress in Materials Science

Received Date: 3 August 2018

Revised Date: 11 October 2019
Accepted Date: 20 November 2019

Please cite this article as: Ramanujam, J., Bishop, D.M., Todorov, T.K., Gunawan, O., Rath, J., Nekovei, R.,
Artegiani, E., Romeo, A., Flexible CIGS, CdTe and a-Si:H based thin film solar cells: A review, Progress in
Materials Science (2019), doi: https://doi.org/10.1016/j.pmatsci.2019.100619

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover
page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version
will undergo additional copyediting, typesetting and review before it is published in its final form, but we are
providing this version to give early visibility of the article. Please note that, during the production process, errors
may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

© 2019 Published by Elsevier Ltd.

 1

Flexible CIGS, CdTe and a-Si:H based thin film solar

cells : A review
Jeyakumar Ramanujam1,*, Douglas M Bishop2, Teodor K Todorov2, Oki Gunawan2, Jatin Rath3*,
Reza Nekovei4, Elisa Artegiani5, Alessandro Romeo6,*
1Materials and Devices Division, CSIR-National Physical Laboratory, New Delhi 110 012, India
2IBM T. J. Watson Research Center, Yorktown Heights, NY, USA
3Utrecht University, Faculty of Science, Debye Institute for Nanomaterials Science-Physics of
Devices, High Tech Campus 5, 5656 AE Eindhoven, The Netherlands
4Department of Electrical Engineering, Texas A&M University-Kingsville, TX 78363, USA
5Laboratory for Photovoltaics and Solid State Physics, Department of Computer Science
University of Verona, Ca' Vignal 2- Strada Le Grazie 15, Verona, VR 37134, Italy
6Laboratory for Applied Physics, Department of Computer Science, University of Verona, Ca'
Vignal 1-Strada Le Grazie 15, Verona, VR 37134, Italy
* Corresponding authors. E-mail: [email protected] (J.R), [email protected] (J.R), [email protected] (A.R)

Abstract

Flexible thin film solar cells such as CIGS, CdTe, and a-Si:H have received worldwide attention.

Until now, Si solar cells dominate the photovoltaic market. Its production cost is a major concern

since Si substrates account for the major cost. One way to reduce the module production cost is to

use the low-cost flexible substrates. It reduces the installation and transportation charges also,

thereby reducing the system price. Apart from metallic foils, plastic films and flexible glass, paper

substrates such as cellulose papers, bank notes, security papers and plain white copying papers are

also used as substrates for flexible solar cells. In this review, recent developments in flexible CIGS,

CdTe and a-Si:H solar cells are reported. Progress on various flexible foils, fabrication and stability

issues, current challenges and solutions to those challenges of using flexible foils, and industrial
 2

scenario are reviewed in detail. Encapsulation issues and solutions related to water vapor

transmission rate are discussed.

1.0 Introduction
Silicon (Si) solar cells dominate the PV market (92%) followed by cadmium telluride

(CdTe, 5%), copper indium gallium selenide (CuInGaSe2 or CIGS, 2%) and amorphous silicon (a-

Si:H, ~1%). Si wafer with thickness around 180 μm is the traditional material being used for

module manufacturing and it has attained significant level of maturity at the industry level. Its

production cost is a major concern for energy applications. About 50% of the cost of Si solar cells

production is due to Si substrate, and device processing and module processing accounts for 20%

and 30% respectively [1].

An alternate to Si solar cells is the thin film solar cells fabricated on glass substrates. The

main demerits of using glass substrates are fragile nature of modules, cost of glass wafer (300-

400μm), and low specific power (kW/kg) etc. Specific power is an important factor when solar

cells are used in space applications. A high specific power exceeding 2kW/kg can be achieved [2]

by flexible solar cells on polymer films which is useful for terrestrial as well as space applications.

Production cost can be lowered by using flexible substrates and roll-to-roll production (R2R)

technique. Apart from light weight, flexibility and less cost of installation, flexible cell processing

involves low thermal budget with low material consumption. Other than solar cell applications,

smaller specialized applications are beginning to become more viable independent markets,

including applications for mobile power and building or product integration, which can benefit

greatly from flexible thin film options. Flexible cells on buildings (known as building integrated

photovoltaics or BIPV) can minimize the cost of support, shipments etc., and installations can be
 3

handled easily. However, flexible solar cell technology is less mature when compared to the cells

fabricated on rigid substrate counterpart.

Due to four main requirements - high efficiency, low-cost production, high throughput and

high specific power, a major research and development focus has been shifted towards flexible

solar cells. It can offer a unique way to reach terawatt scale installation by using high throughput

R2R fabrication technique. Most commonly used substrates are polyimide, polyethylene

terephthalate (PET), polyethylene naphthalate (PEN), and metal foils such as stainless steel (SS)

and titanium (Ti). During fabrication, impurity (such as iron (Fe) and chromium (Cr)) diffusion

from metal foils into the device/active layer is a major concern. Polymer foils are preferable due

to no metallic impurities. It can be used for both substrate and superstrate configurations, whereas

in the case of metallic foils, substrate structure is the only option. Also, due to insulating properties,

polymer foils are suitable for monolithic integration (i.e. cell to cell interconnection). Most

frequently used polymer foils are polyimide. Its maximum processing temperature is limited to

450ºC which is a barrier for high efficiency solar cells, for example CIGS cells. Polyimide foils

are more cost competitive than metal foils. The main concern with superstrate structure by using

polymer foil is the transparency and its stability during high temperature processing conditions.

During the deposition of films, to avoid stress related issues (i.e. delamination, cracking and

poor adhesion) in the film stack, thermal expansion co-efficient of flexible substrate should be in

the range of the deposited films. If the thermal expansion co-efficient of the substrate is too high,

cracking occurs in the film. Also, to circumvent pinholes and shunting, substrate roughness should

be small. Average roughness of 12.5 μm thick polyimide and polished stainless steel are 1.3 nm

and 2 nm respectively [3].

 4

Solar cells on flexible glass have also been reported [4]. As compared to metal foils,

flexible glass benefits from low contamination and roughness; as compared to polymer foils,

flexible glass is compatible with high temperature processing and possesses high optical

transparency which enable more photons to reach the absorber.

CIGS cell efficiency and stability are comparable to Si solar cells. Recently, Tiwari group

(EMPA Switzerland) reported a record efficiency of 20.4% on flexible polymer foil (23% in glass),

independently confirmed by Fraunhofer Institute for Solar Energy Systems Freiburg [5, 6].

Recently, the same group reported an efficiency of 20.8% using low temperature of 450ºC [7].

Similar to CIGS absorber, CdTe is also an excellent absorber material for thin film solar cells due

to its ideal bandgap (~1.5 eV). Close-spaced sublimation (CSS) technique is used to deposit CdTe

films to obtain large grains and high growth rate. On polymer and steel foils, CdTe cell efficiency

is reported to be 13.8% and 10.9% respectively [8]. On flexible willow glass substrate, 16.4%

efficiency (22.1% on rigid glass by First Solar) has been reported [9]. Unlike CIGS and CdTe, a-

Si:H is a disordered material, however, its absorption co-efficient (105/cm) in the visible region of

the solar spectrum is similar to CIGS and CdTe. This is due to disorder in the a-Si:H film and

relaxation rule in the momentum conservation [10]. Flexible a-Si:H solar cells are fabricated either

by direct deposition onto plastic substrates or by a transfer method (i.e. high temperature

deposition on a template and then transfer to a plastic substrate) using Helianthos concept. To

obtain a dense and homogeneous film by plasma enhanced chemical vapor deposition (PECVD,

RF 13.56 MHz) [10], a low deposition rate of 0.1-0.5 nm/sec is used. Hot wire chemical vapor

deposition (HWCVD) is used for high deposition rate and device quality films. For a-Si:H solar

cells, stabilized efficiency of 7.1% [11] and above 10% have been reported on flexible glass and

rigid substrates.
 5

The aim of this paper is to review recent developments in flexible CIGS, CdTe and a-Si:H

solar cells due to their increasing importance. Existing issues to be solved are discussed and

methods to further improve the cell performance are suggested. Future prospects of these flexible

solar cells are outlined. In this review, merits of flexible solar cells are presented in Sec. 2.0.

Flexible CIGS solar cells are reviewed in Sec. 3, while CdTe and a-Si:H solar cells are reviewed

in Sec. 4 and Sec. 5 respectively. These three types are compared and its applications are presented

in Sec. 6. Summary and conclusions are given in Sec. 7.

2.0 Merits of flexible solar cells

The performance of flexible solar cells is comparable to rigid substrates. Flexible substrates

are more advantageous than that of standard soda-lime glass (SLG) substrates. As mentioned

below, there are several merits of using flexible substrates:

 Flexible modules are best suited for curved surfaces and used in BIPV. Since modules are

produced from thin film materials it is suitable for mass production.

 An important benefit is that it has potential to reduce the production cost. R2R deposition

is beneficial in terms of production cost than that of rigid substrates. Glass cover is an

added expense when rigid substrates are used.

 Active materials used for CIGS, CdTe and a-Si:H flexible modules are much cheaper than

conventional Si, glass cover, frames used in Si modules.

 For roof top application, flexible modules are ideal due to light weight. Using lightweight

support, it can be installed over the roof top where glass covered conventional heavy and

bulky Si modules are not suitable when roof test fails due to an added weight and structural

issues. Flexible modules can also be installed over the roof of the vehicle, uneven surfaces

of building.
 6

 Installation/labor cost is much lower for flexible modules due to less installation time since

racking assembly, glass cover etc. are not required.

 Low power output flexible modules for example a-Si:H require large number of modules

to get desired output which can be installed easily above the roof top.

 Glass covered rigid modules are fragile. Flexible modules are not fragile it can be rolled

up, transported and handled easily.

3. 0 Flexible CIGS solar cells

The future of solar cells has often been envisioned to be cost effective R2R manufacturing

of flexible thin film solar cells such as CIGS. To date, however, the vast majority of thin film

manufacturing has taken place with glass-glass encapsulation as a drop in replacement for Si

modules. While the last decade has witnessed extremely rapid growth and progress in the solar

industry, the quickly changing environment and rapidly dropping Si module prices created an

extremely challenging environment for smaller PV companies and less mature PV technologies

which resulted in numerous bankruptcies of thin film aspirants. Nevertheless, CIGS technology

and flexible-production know-how has been continuing to progress with reasonable optimism for

future growth. The dramatic reduction in CIGS module production costs has increased the

importance of balance of systems (BOS) such as module racking, and soft costs such as

engineering and installation time which are key areas where light weight flexible CIGS solar cells

offer advantages. With the growing size of the photovoltaic market, smaller specialized

applications are beginning to become more viable independent markets, including applications for

mobile power and building or product integration, which can benefit greatly from flexible thin film

options. A photograph of CIGS cell on polyimide is shown in Fig.1 (a) while the efficiency trends

for various substrates are shown in Fig.1 (b).

 7

(a)

24
22 Glass
Efficiency (%)

20
18
16
14
12
10
1990 Year 2040
(b)
Fig.1 (a) Photograph showing flexibility of a CIGS device on polyimide with scribed individual cells. Typical structure
consists of substrate/Mo/CIGS/CdS/i-ZnO/Al-ZnO/MgF2, where the notations have their usual meanings. MgF2 is
used to minimize reflection losses. (b) Timeline of record efficiency of CIGS for various substrate types. CIGS on
glass substrates (blue circles) have led the efficiency records with a brief exception in 2013. CIGS on polyimide made
significant gains between 2005 and 2013. The scarcity of SS and ceramic records reflects less research. Figure 1 (a)
reproduced with permission from [12]. Permission granted (No. 4243861044867). Copyright © 2017 Macmillan
Publishers Limited.

Miasole pivoted from rigid glass-glass to flexible SS based modules in 2016 and flexible

thin film tool manufacturers seem to have new momentum in 2017. Despite most CIGS volume

comes from rigid glass-glass panels manufactured by Solar Frontier, a handful of companies have

been offering increasingly efficient flexible CIGS products. CIGS review is split into six sub-

sections, viz. overview of CIGS fabrication and substrates, progress on different substrates,

challenges of flexible substrates, commercial status, fabrication issues and stability, and future

outlook.
 8

Fig. 2 Typical CIGS structure (SLG/Mo/CIGS/CdS/i-ZnO/Al-ZnO/MgF) shown in scanning electron micrograph

obtained from a complete CIGS device on polyimide with efficiency greater than 18%. Here, Mo, CdS, i-ZnO, Al-
ZnO and MgF stands for molybdenum, cadmium sulfide, intrinsic zinc oxide, aluminum doped ZnO and magnesium
fluoride. Reproduced with permission from Ref. [12], Adrian Chirila et al., Highly efficient Cu(In,Ga)Se2 solar cells
grown on flexible polymer films, Nature Materials 10 (2011) 857-861. Permission granted (No. 4243861044867).
Copyright © 2017 Macmillan Publishers Limited.

3.1 Overview of CIGS fabrication and substrates

CIGS, unlike organic and perovskite solar cells, requires processing at high

temperatures (~550˚C) in reactive chalcogen environments to form a high quality absorber layer.

SLG is one of a limited list of low-cost materials that can tolerate such a high temperature and

reactive processing, and it has been a standard substrate for the typical CIGS structure (Fig. 2).

The most common material stack consists of SLG/Mo/CIGS/CdS/i-ZnO/Al-ZnO/MgF followed

by a nickel (Ni)/Al grid. The film is predominantly deposited from the back contact to front

(substrate type) due to the high temperature required for CIGS formation which can destroy the

p-n junction interface if the fabrication order is reversed.

The CIGS layer can be either crystallized in a single step during growth with sufficient

chalcogen vapor pressure (co-evaporation), or formed by depositing a precursor film (typically a

metal stack), which is then reacted by annealing in a sulfur/selenium environment. The numerous

vacuum (evaporation, sputtering, etc.) or non-vacuum (solution, nanoparticle) methods are

discussed in detail elsewhere [13-15]. Traditionally, the highest efficiency devices have been
 9

formed by co-evaporation due to the increased freedom to tailor the film composition and

particularly [Ga]/[group III] ratio throughout the film depth. This allows for controlled bandgap

grading, which has been found to be critical for high efficiency cells. In recent years, however,

Solar Frontier has challenged this orthodoxy and demonstrated a number of record cells (including

22.8% uncertified [16] using a sputtering and selenization based method. For these record cells,

the bandgap grading profile is achieved through multi-step selenization with a final stage using

sulfur.

Incorporation of alkali elements, such as sodium, potassium and cesium, is performed to

passivate the CIGS layer (among other effects), and has become increasingly common as a post-

deposition treatment after the CIGS layer is formed. Sodium incorporation has long been known

to be beneficial to CIGS performance [17, 18], however, in recent years it has become clear that

potassium and heavier alkali elements (cesium and rubidium) are essential to improve the

performance beyond 20%. Since potassium was first applied to the record 20.4% cell from EMPA,

multiple subsequent efficiency records have been achieved by implementing and optimizing

potassium post-deposition treatments. The most recent 22.9% record was enabled by optimizing

post-deposition treatments with heavier alkali elements [19]. For flexible cells, the treatments are

essential because no Na is provided by the base substrate as is the case with SLG (although it is

worth noting that most high efficiency cells on glass, as well as industry, rely on controlled NaF

additions and often block diffusion from SLG).

The traditional n-type buffer layer in high-performance research cells is CdS typically

deposited by chemical bath deposition, but industry often employs vacuum deposition of

Zn(O,S,OH) and or (Zn,Mg)O. This not only reduces the environmental hazards associated with

Cd usage but also allows higher current due to more transmission in the range 350-500 nm. Non-
 10

CdS buffer layers have been employed with most success by Solar Frontier who has achieved an

uncertified efficiency of 22.8% using a Zn-based buffer layer [16, 20]. The final layers including

an intrinsic ZnO layer and a transparent conductive oxide (TCO, viz. ZnO:Al) are often deposited

by sputtering which are followed with Ni/Al grids which can be either screen printed or deposited

through vacuum techniques. The intrinsic ZnO is used to prevent shunting paths caused by the

diffusion of Al from ZnO:Al into the absorber and conductive ZnO (i.e. ZnO:Al) is used for the

lateral collection of charge carriers into the top current collecting grids [6]. The high temperature

annealing or growth step still dictates many of the requirements of the fabrication sequence

including substrate choice which are discussed below.

Table 1: Recent record CIGS devices for various substrate types (* uncertified efficiency).
Efficiency in glass substrates are given for comparison.

Substrate Eff. (%) Voc Jsc FF Year Company/ Alkali metals/ Ref.
(mV) 2
(mA/cm ) (%) Institution buffer layer

Glass 22.9 746 38.5 79.7 2018 Solar Cesium [19]

Frontier

22.6 741 37.8 80.6 2016 ZSW KF + RbF [21]

22.8* 711 41.4 77.5 2016 Solar KF, with ZnMgO [16]
Frontier Zn(OS, OH)
buffer

20.8 757 35.8 79.1 2014 ZSW KF [22]

Polyimide 20.4 736 35.1 78.9 2013 EMPA KF [23]

SS 19.4* 760 32.6 79

(50 mil) 19.1 750 33.0 77 2017 Miasole KF [23]

SS 18.7* 720 35 74.4 2017 Global KF [25]

(25 mil) Solar
 11

3.2 Progress on different substrates

Cell efficiency and other parameters for recent champion cells for various substrate types

are listed in Table 1. The champion cells are fabricated on polyimide and SS substrates [23-25].

On polyimide, high efficiency has been reported to be 20.4% with Voc of 736 mV, Jsc of 35.1

mA/cm2 and FF of 78.9%. On SS, the corresponding values are 19.4%, 760 mV, 32.6 mA/cm2 and

79%. For about last two decades, flexible substrates have lagged the efficiency records of CIGS

on glass as shown in Fig. 1(b). CIGS on polyimide has historically been only 2/3 the efficiency of

glass champion cells, however, in recent years significant progress has been made, and in 2013,

flexible polyimide was for a time, more efficient than glass counterparts. The relative frequency

of CIGS records on glass and the more spaced timing between polyimide and stainless steel records

also reflect the relative research volumes for each substrate. CIGS records depicted in Fig.1(b)

have been set by 5 institutions over the last 5 years, while record polyimide devices have been set

7 times by a single institution (EMPA) [23, 26]. Similarly, few institutions (EMPA, Miasole and

Global Solar) have been intensively optimizing SS substrates and fewer still have been optimizing

on flexible ceramics (AIST) or other metal foils which is reflected in the scarcity of record data

points. Despite this, back in 2008, Ti foil and flexible ceramics based on Zirconium Oxide (ZrO2)

achieved 17.4% and 17.9% respectively suggesting strong efficiency potential for these substrates

[27].

As noted by the rapid progress in polyimide in the years between 2005 and 2013, the

current state of records does not necessarily reflect technological limitation of each substrate – but

rather the current state of optimization. The technical barriers for flexible substrates are not

unsolvable. It is likely that today’s efficiency gap between flexible substrates (20.8%) and the

record 22.9% [7, 19] cell can be narrowed significantly by applying recent lessons pioneered on
 12

glass such as optimizing heavy alkali earth elements and improving buffer layers. Therefore, the

future progress of flexible CIGS is bright.

3.3 Challenges of flexible substrates

As mentioned above, flexible substrate choice includes metal foils (most commonly SS),

polyimide sheets, and thin flexible ceramics. The default fabrication recipes, optimized around

SLG properties, require tweaking for application to flexible substrates.

Flexible CIGS cells allow monolithic integration and offer high output voltage and low

production cost. Three requirements of photovoltaic modules are low production cost of cells, high

efficiency and easy installation. To fulfill these requirements, flexible substrates should have high

thermal stability, thermal expansion co-efficient matching with CIGS for good adhesion, excellent

dielectric properties, and no impurity diffusion from substrates. Flexible substrates do not satisfy

all these requirements together. For example, Al is an excellent low cost material for thermal

stability, however its thermal expansion co-efficient does not match with that of CIGS film [28].

Ceramic foil such as ZrO2 can meet the above requirements, nevertheless due to its high cost and

brittle nature it is not suitable for R2R production.

Polyimide substrates, which are attractive due to their light weight, and highly flexible

nature, suffer primarily from poor high temperature stability which limits fabrication temperature

to around 450˚C. Low temperature fabrication recipes have been attempted for many years but

efficiencies have lagged and remained below 15% until 2000 and still below 16% since 2010 [12,

29]. The efficiency record for polyimide is 20.4%, obtained by using low temperature three-stage

deposition process followed by a sodium fluoride (NaF) and potassium fluoride (KF) post-

deposition treatment [23]. In 2019, an improved efficiency of 20.8% has been reported [7]. The

co-evaporation process is amenable to low temperature due to minimized elemental diffusion

 13

distances which require time and temperature as well as the ability to directly tailor the bandgap

and elemental grading through the film depth.

Due to the lack of a Na source inherent in flexible substrates, alkali (NaF, KF, and

Rubidium fluoride RbF at <400ºC) addition timing – before, during or after the CIGS growth

process – have been found to be critical to device efficiency [21, 30, 31].

Unlike polyimide, SS can withstand the high temperatures used for SLG, however,

diffusion of impurities (Fe, Ni, Cr) from the steel substrate, particularly Fe, forms deep states in

the CIGS absorber and therefore can limit the efficiency [32]. Diffusion barriers have been often

used to combat this challenge including the use of Al-oxide (Al2O3), Si3N4, Ti-nitride (TiN), nickel

phosphorus (NiP) [33], or enameled steel which adds an extra cost, however, tailoring the Mo back

contact and reducing fabrication temperature have also been shown to be successful methods to

mitigate this problem.

Thermal expansion co-efficient of SS match with that of CIGS film [34]. In 2011,

EMPA demonstrated a 17.7% CIGS cell on stainless steel without a diffusion barrier by increasing

molybdenum density and thickness and transferring the low temperature co-evaporation processes

used for polyimide [35].

The conductivity of SS or other metallic substrates creates an additional challenge for

monolithic integration of multiple cells on a single substrate. For monolithic integration, flexible

substrates should have insulating property or an insulating layer must be added between the

substrate and cell Mo back contact in order to allow isolation between cells. This layer often serves

dual purpose as the diffusion barrier for impurities. Low cost technique is not available for

insulating layer deposition with good insulating properties. As an alternative to monolithic

integration, cells can be cut and tabbed or wired, as it is done for Si solar cells. This strategy, while
 14

less elegant, provides the advantage that cells can be binned with common efficiency to minimize

cell bottlenecks that could be present in a full module.

A final concern for flexible substrates is flexible encapsulation in order to provide a strong

environmental barrier for the 20 years life-span expected for solar cells. More details of flexible

CIGS and encapsulation are reviewed in Ref. [26].

3.4 Commercial status

CIGS has a long history of research and commercialization. A thorough history and

retrospective look at CIGS was recently published [36]. Industrial work on flexible CIGS dates

back to ISET in 1992 [26]. However, as previously mentioned, glass substrates have provided a

more mature production process, and today the vast majority of CIGS product manufactured has

a rigid glass-glass encapsulated structure, a large fraction of which is produced by Solar Frontier

(~1GW in 2016) [37].

Average selling price for CIGS PV modules which had been stable at approximately

$4/Watt in the years leading up to 2007, plummeted to less than $1/Watt 5 years later, and have

continued to decline to less than $0.5/Watt by the end of 2016. Multiple leading manufacturers in

the first quarter of 2017 claim production costs [37] of less than $0.40/Watt (excluding BOS and

installation charges). Production cost can be lowered further by using flexible substrates.

Today, more than a dozen companies continue to make progress in the CIGS technology,

and a number of notable examples are demonstrating the potential of flexible CIGS. An incomplete

list of current companies producing flexible solar cells is shown in Table 2. The table also includes

notable CIGS companies that have shut down in the last few years. A more complete list and details

of industrial CIGS players can be found elsewhere [13, 26].

 15

Table 2: A list of flexible CIGS companies, past and present.

Company CIGS Status Notable results

deposition on
flexible
substrates

Miasole Sputtering Purchased by Hanergy in 19.4% cell, 17%

(Hanergy) Stainless steel 2013 app., module
substrate

Global Solar Co-evaporation Purchased by Hanergy in 18.7% cell

(Hanergy) Stainless steel 2013

Flisom Co-evaporation Equipment manufacturing 16.6% small sub-

Polyimide partners with EMPA with module
20.4% record polyimide

Midsummer Sputtering Equipment manufacturer 18.8% aperture area

Stainless steel with multiple order in 2017 6in cell (16.4% total
substrate area)
Ascent Sputtering Purchased by TFG Radiant -
Polyimide in 2011, shifted to solar
charging market. Currently
restructuring loans to
continue operations

Nuvosun Sputtering Purchased by Dow – CIGS -

(Dow) Stainless steel BIPV roof tile but no longer
substrate being manufactured – now
licensing silicon based
shingle to RGS

Solarion - Purchased in 2015 – now -

Polyimide selling Si modules
substrate

Nanosolar Nanoparticle ink Bankrupt 2013 17.1% cell

Al substrate

Sputtering
Odersun Cu substrate Bankrupt 2012 -

Electrodeposition Bankrupt 2013, minimal -

Solopower Stainless steel operation under name
substrate SoloPower Systems as of A
 16

A number of notable flexible CIGS results are worth highlighting. Miasole, which pivoted

recently to all stainless steel modules in 2016 reported a record efficiency of 19.4% for a small

cell and module aperture area efficiency of ~17% [24]. Global Solar uses a slightly thinner (1 mil)

SS substrate has also made significant progress by demonstrating a champion cell of 17.7% in

2016 [38], and 18.7% in 2017 [25]. Finally, Flisom is commercializing equipment to manufacture

CIGS on flexible polyimide in collaboration with EMPA.

Although sales volume is a small fraction of Si, with production lines measured in tens of

MW, flexible CIGS product maturity and sales volume appear to be growing in recent years.

Flexible CIGS tool manufacturers such as Midsummer also showed momentum in selling their

R2R CIGS on SS with multiple product orders in 2017. Even Solar Frontier, the leading CIGS

manufacturer by volume, demonstrated SS based flexible CIGS panels in 2015.

3.5 Fabrication issues and stability

Despite the promises, flexible CIGS creates a number of fabrication issues for

manufacturers. Processing conditions established for an acceptable compositional grading profile

for CIGS is not easy to maintain during production. When SLG substrates are used, during CIGS

deposition, sodium is supplied by the substrate. For flexible substrates, sodium containing film is

required prior to back contact (Mo) deposition. Sodium doping in CIGS (~1019/cm3) [39] provides

passivation and improves electrical performance and yields high Voc and fill factor. When sodium

free substrates (i.e. flexible substrates) are used, accurate level of Na doping in CIGS from sodium

containing film is an issue. As a side note, doping level depends on the thickness of sodium

containing layer.

When flexible substrates are used, Ga grading in CIGS is steeper than that of rigid (SLG)

substrate for similar processing steps. This is due to the variation in indium and Ga diffusion rates
 17

[28]. Steeper Ga grading results in narrow band gap CIGS that results in lower Voc and higher Jsc.

Another issue is laser patterning when flexible substrates are used. When the substrates are thin,

process integration is critical since laser beam focusing is difficult [39] which increases leakage

current in the device. Processing conditions used for thick and SLG substrates cannot be used for

flexible substrates.

Beyond the typical issues of producing CIGS at scale, the flexible substrates must

withstand the high temperatures and reactive chalcogenide environments used for CIGS

fabrication. For polyimide substrates, this places an upper limit on the fabrication temperatures

which traditionally have been used to achieve high efficiency. Some metal foils, such as SS can

withstand the high temperatures but can be limited by impurity diffusion of elements such as Fe

from the substrate into the CIGS layer. A metal foil or conductive substrates in general, are also

less advantages to monolithic integration, requiring either an extra insulating layer or an alternative

interconnection scheme. Other substrates such as flexible ceramics or flexible glass without such

problems have also been explored, however, there remain open questions about substrate cost, and

the vast majority of research as well as the history of champion CIGS efficiencies has continued

to be on traditional SLG. Due to these issues, the concept of a 20% efficient flexible CIGS module

had seemed out of reach until recently. In 2013, EMPA (Switzerland) demonstrated a record

efficient CIGS cell on polyimide, surpassing even the glass-based CIGS of the time [23]. This

demonstrates that despite processing challenges, bridging the efficiency gap between rigid and

flexible cells is feasible.

Mechanical robustness of flexible CIGS cells is affected by stress in the film (internal

stress) and stress induced by bending the module. The latter depends on the film thickness,

Young’s modulus of the substrate and film. Strain in the CIGS film induced by bending (radius 50
 18

mm) test is higher in ultra-thin (100 µm) flexible glass substrates than that of 25 µm thick

polyimide substrates. Bending induced degradation of solar cells is irreversible and deterioration

is permanent when reset into initial state [40]. More deterioration was noticed in the most bent

region, mainly due to a slight increase in series resistance which decreases fill factor, Voc, Jsc.

Degradation can be lowered by using substrates having low thickness and Young’s modulus.

Metal foil is used for high mechanical and thermal stability (>600°C) [41]. However, it

requires diffusion barrier as mentioned as above. Commercial metal sheet contain deep grooves,

spikes and cavities over the surface since foils are rolled up during manufacturing, and undergo

bending. To improve mechanical robustness of the device, foils are polished or a layer is coated to

level the foil. A good mechanical stability has been reported for flexible CIGS cells [12].

Bending induces stress in the module leads to cracks or delamination. A strong adhesion

between back contact (Mo) and CIGS is required to avoid stress [29]. During high temperature

(>450°C) deposition of CIGS film, thermal expansion co-efficient (heat expansion) of flexible

glass should be close to that of CIGS film. Otherwise adhesion issues and cracks can occur. For

ultra-thin glass, Mo and CIGS, TEC is around 5-12x10-6/K [29, 40].

In flexible cell, unbearable external loads during the operation can produce internal stress

leading to cracks and interface layers split apart [42]. For flexible CIGS cell, a tolerable limit for

compressive stress without cracks is around 70 MPa [43].

Long term environmental stability is one of the main factors for commercialization. CIGS

films are tolerant against high energy radiation. Unlike other technologies like perovskite solar

cells, CIGS stability is not an issue. However, one of the main concerns is the protection from

environment and flexible encapsulation with a life-span of 20 years.

 19

3.6 Future outlook

CIGS solar cells offer the highest efficiency and mature flexible solar cells for mainstream

applications. The efficiency outmatches alternatives such as dye-sensitized solar cells and organic

solar cells, and unlike perovskites, stability is not an obstacle, and toxicity concerns are minor or

surmountable. In fact, flexible CIGS modules shipping today now rival the efficiency of multi-

crystalline Si solar cells. It is not yet clear whether the future of flexible CIGS will stay positioned

in niches where flexibility is integral to the product, or it will vie for larger portion of the solar

market. The past decade, which has been competitive for solar cell manufacturers have left a

stronger group of pioneers that may give flexible CIGS a tough chance.

4.0 Flexible CdTe solar cells

Among thin films, CdTe has demonstrated the highest scalability and reproducibility so

far, as reported by the successful industrial production within the last 10 years. The high absorption

coefficient allows producing high efficient devices also with ultra-thin absorbers, values exceeding

10% have been obtained with thickness below 1 μm [44, 45] and around 10% for only 0.5 μm [46].

The CdTe thickness reduction could be a crucial point for increasing efficiency and stability of

flexible devices. Both rigid and flexible CdTe solar cells perform best when made in superstrate

configuration, and this happens mainly for two reasons:

(i) The special process required with the so-called “activation treatment” where typically a Cd

chloride (CdCl2) thin layer is deposited on top of the CdTe and then the stack is annealed in air

[47, 48]; the activation treatment acts as a step for increasing the electrical properties of the

absorber and for improving the junction between CdS and CdTe [49]. So if it is made after CdTe

deposition on CdS, the performance is higher.

 20

(ii) The back contact generally requires an addition of Cu for high performance. In substrate

configuration Cu is deposited on the back of the solar cell, and during CdTe deposition (with high

substrate temperature) Cu reacts with CdTe film, which causes degradation in the film [50-52].

Fig.3 Schematic of superstrate and substrate CdTe devices (the substrates’ width is not proportional but indicative of
the differences between the devices). Both in substrate and superstrate configurations, light enters through wide
bandgap window material (CdS). In the substrate structure, opaque substrates are used, CdS need not undergo high
temperature processing.

Recent studies at EMPA have shown that in substrate configuration it is possible to reduce

these limitations by depositing a Cu-free back contact, namely MoOx, followed by CdTe and an

ultra-thin Cu layer on CdTe; moreover, CdCl2 is applied before and after CdS deposition. Cu

diffusion into CdTe is performed at around 400˚C. A record efficiency for CdTe substrate cells of

13.6% has been obtained [53]. In this way, the temperature and time of annealing strictly control

Cu diffusion; on the other hand, if Cu is deposited prior to CdTe deposition, this diffusion will

depend on external parameters since the time at which the Cu is kept at high temperature is

depending on the CdTe deposition rate.

The CdCl2 treatment promotes recrystallization of the CdTe and enhances grains growth;

it also improves the CdTe/CdS junction properties by enhancing inter-diffusion of sulphur into the
 21

CdTe and reducing the lattice mismatch between the two layers. So the treatment has a double

purpose, which makes it necessary to be used after the CdTe deposition. When CdTe is deposited

in substrate configuration, CdCl2 has to be applied two times, a first time before and a second time

after CdS deposition [53, 54].

CdTe solar cells in superstrate configuration have reached the superior efficiency of 22.1%,

reported by First Solar Inc. So superstrate configuration is still preferable also for flexible devices.

On the other hand, this requires special substrates that have to fulfill transparency and flexibility

requirements at the same time. Various configurations of CdTe solar cells are shown in Fig.3.

It is also very important to mention two additional improvements that have been done

recently to increase efficiency that are the CdS substitution with a more transparent buffer layer,

(i) magnesium zinc oxide is a very good solution [55] and (ii) CdSexTe1-x intermediate layer for

band gap grading [56-58].

4.1 State of the art: Fabrication methods and issues

The paths for fabricating efficient CdTe solar cells have been divided into two main cases:

(i) To optimize a fabrication process for substrate configuration CdTe solar cells by using a non-

transparent substrate such as metal-based or polymer-based layers, and (ii) To fabricate and

optimize a flexible CdTe cells on transparent substrate capable of withstanding high process

temperatures. In Fig.4, CdTe cells on rigid glass, flexible cells on ultra-thin glass (such as willow

glass) and polyimide substrate are shown.

 22

Fig.4. From left to right: CdTe cells on glass, ultra-thin glass and polyimide made at the laboratory for photovoltaic
and solid state physics, University of Verona. The cells on polyimide are more flexible than the cells deposited on
ultra-thin glass. However, polymer results in a reduced transparency compared to ultra-thin glass, as it is clearly visible
in the picture.

4.1.1 Substrate configuration cells

In the case (i), the most commonly used substrates are Mo or thick polyimide (for example

Kapton®). In any case best working devices have been fabricated using Mo as back contact, either

deposited on a Kapton® or a Mo sheet.

Mo allows a superior stability on the subsequent depositions, and at the same time, it is a

good conductive metal for electrical contact. However, it does not have high enough work function

for providing a good Ohmic contact. On the contrary, a small energy barrier is formed if pure Mo

is contacted with CdTe. This has been proved by Williams et al. [54], where a barrier height of

0.51 eV has been reported. In order to overcome or at least limit this barrier issue, the following

two steps can be introduced:

 Deposition of an interfacial layer with high work function between Mo and CdTe.

 Increase CdTe doping on the backside (close to back contact) of the CdTe layer.

Dhere et al. have studied different interfacial layer for CdTe substrate configuration

devices, where CdTe is deposited by close-space sublimation at high substrate temperatures [59].

A CuxTe layer is deposited on top of Mo prior to the CdTe deposition, this might not only improve

the back contact performance but also enhances CdTe doping by forming a p+ zone near the back

contact. In Ref. [60], it has been shown that MoOx deposited by RF sputtering from a compound
 23

semiconductor target allows a good performing device with suitable back contact. However two

more steps are crucial for a good substrate type solar cell:

 Double CdCl2 annealing (on the CdTe and CdS/CdTe junction)

 CdTe doping

Kranz et al. have brilliantly solved these issues by applying CdCl2 treatment after CdTe

deposition and after CdS deposition, moreover CdTe have been doped by depositing a monolayer

of Cu on the CdTe/MoOx/Mo stack and annealing [53]. In this way, a record efficiency for

substrate cells of 12.6% on glass substrate [53] has been obtained.

4.1.2 Superstrate configuration cells on flexible and transparent substrates

In the case (ii), the main possibility for producing flexible CdTe-based devices is to

maintain the most productive superstrate configuration and study an innovative highly temperature

resistant substrate, which is at the same time transparent enough to allow UV/visible radiation to

be absorbed by the CdTe layer. In order to fabricate a superstrate solar cell, the substrate has to be

stable under the different processes that are applied for the device fabrication. The substrate has to

withstand in the process sequence: TCO deposition, CdS and CdTe deposition, CdCl2 activation

treatment, back contact deposition and annealing. In particular for high efficiency cells, CdTe is

deposited by high substrate temperature processes such as CSS or vapor transport deposition. For

above 500°C, it is extremely difficult to find a suitable substrate. For this reason, historically, using

low temperature CdTe process [2], the first superstrate type CdTe solar cells have been made on a

soft polyimide foil.

Different polymers can be used for cell fabrication but the best ones are Upilex-S by UBE

and CPI (Crystal Clear) Kapton® by Du Pont [61, 62]. The latter performs a remarkable

transparency also at the low wavelength region allowing higher photocurrents. However, the
 24

processing temperature cannot go over 450°C, so the most used technique is thermal evaporation.

The best efficiency of 13.8% has been reached so far by thermal evaporation of CdS and CdTe on

a ZnO:Al/ZnO stack [62].

More recently, a new type of substrate has been available, which overcomes the limitations

of polymers - a flexible glass, with a thickness of 100 μm. In this case, an alkali free glass with a

thickness of 100 μm becomes perfectly flexible, but at the same time remains stable under high

temperatures and chemical treatments.

Table 3: Best efficiencies of flexible CdTe solar cells

Structure Substrate Lab Substrate Eff. (%) Voc Jsc FF (%) Ref.
T (°C) (mV) (mA/cm2)

Superstrate Flex glass NREL >550 16.4% 831 25.5 77.4 [9]

PI film EMPA <450 13.6% 846 22.3 73.4 [63]

Substrate Mo foil EMPA <450 11.5% 821 21.8 63.9 [53]

Solar cells based on flexible glass can be typically prepared with high substrate temperature

processes. The first introduction of ultra-thin flexible Corning glass for CdTe devices has been

made by Rance et al. [64]. The devices were made by close space sublimation at temperatures

above 600°C, after CdCl2 vapor treatment and deposition of ZnTe:Cu back contact, a record

efficiency of 14.05% was obtained. This value has been overcome by researchers from the same

group, by the application of a sputtered CdS:O instead of the prior chemical bath deposited CdS,

gaining in transparency which resulted in a higher response in the blue light region with a higher

efficiency of 16.4% [9]. A summary of the best efficiencies for CdTe flexible solar cells is

presented in Table 3.
 25

4.2 Issues in flexible CdTe solar cells

Solar modules require long-term stability (~20 years). This is even a more important issue

when it comes to flexible devices where there must be no change in efficiency as the module is

being flexed. Stability and fabrication issues are discussed below.

4.2.1 Stability issues

CdTe solar cells, whether built in superstrate or in substrate configuration, have to perform

long-term stability. Generally CdTe is a very robust and stable material and no particular problems

have been addressed for stability when suitable encapsulation against moisture is applied. In earlier

times an age-old issue of performance stability was due to copper diffusion. However this has been

practically solved by strongly reducing the copper amount, and by adding a copper barrier at the

back contact, such as ZnTe that acts as a barrier to Cu [65] as also previously mentioned. Other

solutions to this problem also have been introduced such as a barrier or As2Te3 [66] or the

incorporation of copper in the form of a chlorine salt [67], which allows introducing an extremely

low amount of copper by combining it with the CdTe back surface. However now CdTe based

solar cells on glass have demonstrated an optimal stability for long time performance as also

demonstrated by the large module production.

For flexible solar cells, new issues have to be considered in terms of stability. The bending

and stretching of the module should not affect the performance of the device. Only few reports are

available for the stability of flexible devices. Rance et al. show a study of the effects of tensile and

compressive stresses for flexible cells on ultra-thin glass [64]. The cells were fastened with PVC

tubes in order to be able to bend them. Different curvatures were experimented and tested and the

resulting efficiencies were measured. Analogous data for cells on polymers were not reported,

however, the change in efficiencies (in some cases efficiency increases when cells are bended) are
 26

not attributed to the type of substrate but to the stress in the CdTe itself, so similar results would

be expected for the cells on polymer substrates [64].

Tensile and compressive stresses have been measured by applying Raman spectroscopy on

CdTe devices deposited on both polyimides (CPI and Upilex-S) and ultra-thin glass (Schott) based

substrates [68]. The data obtained are coherent with the results reported by Rance et al. [64], and

the conclusions are that polyimide substrates give a higher distortion to the CdTe layer compared

to ultra-thin flexible glass. Rigid glass and ultra-thin glass have very similar thermal expansion

coefficients.

Moreover, in the case of cells having polyimide substrates, a very important issue is the

stability of the substrate to ultra-violet (UV) radiation. In this case, an anti-UV coating has to be

deposited on the other side of the substrate.

4.2.2 Module structure fabrication challenges

Polyimide based cells require an optimization of laser scribing technique to avoid substrate

damage. A specific analysis for interconnection of polyimide based CdTe mini-modules has been

presented in Ref. [62], where they demonstrated the possibility of using a conventional laser

scribing technique for the fabrication of a mini-module with 9.4% efficiency on 31.9 cm2 active

area (i.e. area not covered by the grid lines), confirming the feasibility of flexible CdTe module

production even with polymer substrates.

In terms of ultra-thin glass no issues on the laser scribing process occurs, and a very similar

approach to the rigid substrate can be used. So with a smart management of the in-line production

on thin glass (breaking of the substrate has to be avoided), especially in terms of avoiding

temperature shocks by designing proper temperature ramps, fabrication of flexible modules on

 27

ultra-thin glass would just use the same technology which is already widely applied in mass

production.

On the other hand, for both polyimide and thin glass based cells, a crucial point in terms of

scalability is the possibility of a R2R production that would allow high feasibility for large scale

fabrication. For this reason, again it is important that bending of the stack should not affect the

performance of the cells. We have already demonstrated in the above section that no particular

problems occur for ultra-thin glass, while for polymers some larger tensile stresses might affect

the initial efficiency.

Moreover, while ultra-thin glass allows the same transparency as a solid device, for

polymers the absorption at low wavelength regions is not negligible. Most of the polymers (such

as Kapton® and UBE-Upilex S) require an extremely reduced thickness, typically 7.5µm [62, 68].

On the other hand, improved Kapton® polymers such as CPI show a superior transparency and are

rather independent from the thickness [63].

4.3 Future developments

Contrarily to other thin film materials, CdTe solar cells have come to large mass

production, for the moment concentrated on one company. The simple stoichiometry and high

reproducibility reduce problems with inhomogeneity, which are more frequent for other high

efficiency thin film solar cells. However, the large mass production of Si wafers has put production

costs of thin film modules just at the same level of Si modules.

Actually thin films have a higher potential for cost reduction if we can take the advantage

of their ability to adapt to different shapes. This will allow integrating and substituting

conventional roofs with energy producing roofs.

 28

Application of CdTe to new substrates must be studied since CdTe is an extremely simple

material to be grown and it is not affected by the different substrates, has reduced issues in

inhomogeneity for large scale, it has a large potential of mass production in flexible configuration.

Superstrate devices are the easiest to be fabricated, they would be optimal for windows and

flexible glass while substrate devices would be preferred on tiles and ceramic substrates.

Improvements for flexible CdTe photovoltaic devices in substrate configuration should include:

 Development of innovative substrates that can be combined with materials for buildings.

 Introduction of new buffer layers to protect devices from impurity diffusion from the

substrate.

 Application of new back contacts in order to have efficiency improvement and stability to

the subsequent deposition processes.

 Implementation of improved doping techniques in order to enhance electrical properties.

Improvements for flexible CdTe photovoltaic devices in superstrate configuration should include:

 Development of low-cost ultra-thin glass and polymer substrates.

 Application of UV protection layer for polymer films.

In both configurations a thorough study of degradation and stability during bending of the cells

would be required.

5. Flexible a-Si:H solar cells

The success of a-Si:H thin film materials is based on the high throughput and large area

deposition processes for solar cells and display technology applications. Though several types of

inorganic thin films (GaAs, CIGS, CdTe etc.) have been reported [69], a-Si:H type of cells has

shown the widest range of products among inorganic thin films, e.g., a-Si:H solar modules on glass

substrate of 5.7 m2 size or long rollable flexible substrates. This is possible due to the CVD
 29

deposition processes, especially the very high frequency (VHF) PECVD that allows a easy scaling

up or movable substrate on large dimensions and long depositions without down time. The

depositions on flexible substrates is especially notable, ranging from metals (SS, Al) to high

temperature plastics (polyimide or Kapton®) to low temperature or temperature sensitive plastics

such as polycarbonate (PC), polyethersulfone (PES), PET, PEN and papers (cellulose papers, bank

notes and security papers). Depending on the transparency of the substrate, substrate type (n-i-p)

and superstrate type (p-i-n) cells are made. Direct fabrication with n-i-p configuration is made on

metal foils, high temperature plastics (polyimide or Kapton®), or any opaque materials like tiles,

paper etc., whereas low temperature plastics (PC, PET, PEN) and even thin flexible glass mostly

use p-i-n configuration. A lift-off technique (transfer type) however can be used on any type of

substrate. The status of flexible a-Si:H solar cells, and their long term potential as energy source

are reviewed below.

5.1 Flexible substrates: Metal foils, plastics, papers and flexible glass
As mentioned above, flexible cells are mainly categorized into two types based on the type

of substrate: (i) metal foils and (ii) plastics and papers. A third category, based on thin glass is also

worth noting. A review on them is presented below.

5.1.1 Metal foils

Both SS and Al foils have been used for a-Si:H solar cells. The former is well established

commercially, especially by United Solar Ovonic (ECD) known as Uni-Solar. One of the main

disadvantages of these types of cells is the difficulty in monolithic integration that is hampered by

an opaque, in addition to a conducting substrate. Substrate type single junction a-Si:H, double

junction a-Si:H/a-SiGe:H and triple junctions a-Si:H/a-SiGe:H/a-SiGe:H, a-Si:H/a-SiGe:H/nc-

Si:H and a-Si:H/nc-Si:H/nc-Si:H solar cells with PECVD deposition of semiconductor layers have
 30

been reported. The SS foil needs a planarization step to mitigate the effect of spikes or protruding

features. One of the ways is to use UV lacquer that serves two additional purposes: (i) the

electrically insulating layer on steel foil enables monolithic series interconnection [70], (ii)

nanoimprint lithography can be done to enhance light scattering effect [71]. Anodization and

nanoprint lithography have emerged as two main techniques for fabricating textured surfaces. Al

substrate allows texturization by anodization [72] and nanopatterned plasmonic texture can be

made [73]. R2R anodized electrodeposited Ni-Fe foil based a-Si:H cells are also used for water

splitting [74]. Nanodent textures by anodization of Ti foil is another option [75].

5.1.2 Plastics and papers

There are basically two types of substrates: expensive plastic substrates for high

temperature processing and low-cost temperature sensitive substrates for flexible a-Si:H solar

cells. These two types are mainly distinguished by their glass transition temperature (Tg).

Fabrication on these substrates are reviewed and presented below.

(a) High temperature resistance substrates: These expensive plastics whose Tg is higher than

that of the thin film Si materials allow the semiconductor layers to be deposited at their optimum

temperature condition (~200ºC). These plastics also show resistance to the other layer depositions

(silver, TCO etc.). Polyimide (Kapton® from DuPont) has been extensively used, though

commercially is not very attractive due to high cost. Powerfilm (Iowa thin film) and Fuji Electric

make such modules. For these opaque substrates, the monolithic series integration by R2R

processed “series connection through apertures formed on film” (SCAF) structure was used by

Fuji electric. An efficiency of ~12% for a-Si:H/nc-Si:H solar cell in n-i-p configuration has been

achieved.

(b) Low-cost substrates: In these temperature sensitive substrates, their Tg is well below the Tg

of the Si layers. PEN, PET, PC, PES and various types of “papers” are in this category. These
 31

substrates have received lots of interest, as they not only have the potential to reduce the cost, but

also some of them allow to fabricate the device in superstrate configuration, which has definite

advantage as far as monolith integration for module manufacturing is concerned. This type of

substrate is challenging to fabricate the semiconducting (Si intrinsic and doped) layer and the

textured front TCO on them for superstrate type cell or textured back reflector in substrate type

cell, which are made optimally at temperature much higher than the substrate deformation. Laser

scribing also poses difficulties, as the temperature can shoot up locally. Moreover, the post

fabrication annealing and lamination are issues. Hence, two methods are followed to address these

issues. The first is the fabrication of complete cell on a temperature resistant temporary (sacrificial)

substrate, thus, the layers deposited at their optimum temperature, and then transferred to a

permanent substrate, and the second is to find parameter space to grow layers at temperatures

below 150ºC. These two cases are discussed below.

(i) Transfer type (lift-off) process: There are various transfer methods developed for all sorts of

inorganic type flexible solar cells [76]. Helianthos concept is the best example of this type for a-

Si:H cell. An a-Si:H cell is made on a metal (kitchen Al) foil with the following sequence:

deposition of textured TCO (Fluorine-doped tin oxide, FTO) at 400-500ºC by atmospheric pressure

chemical vapor deposition (APCVD) using tin tetrachloride as precursor, a-Si:H p-i-n layers at

150-200ºC by PECVD, TCO (ZnO:Al) by PVD (magnetron sputtering), Ag metal contact by PVD

(evaporation or sputtering), laser scribing, and then attach it to a permanent substrate (PET),

followed by removal of Al foil by sodium hydroxide, lamination of the front side and the backside

with low water vapor transmission rate (WVTR) polymer encapsulant. The whole process is done

in R2R production of PVD and PECVD machines. This p-i-n processing is extended to a-Si:H/nc-

Si:H tandem cell. Akzo-Nobel along with Utrecht University invented this cell processing, for
 32

which subsequently TUD, TUE, ECN and TNO from Netherlands and FZ Juelich from Germany

contributed significantly to develop to module stage. In the standard PECVD process, 35 cm width

foil is used by the industry for the pilot module production and further development is needed. As

the belt size in R2R process depends on deposition rate, a high rate deposition process is desirable

especially for the tandem cell manufacturing that needs a thick nc-Si:H (~1500 nm) bottom cell.

VHF PECVD was explored to this end and an efficiency of 7.9% for single junction and 8.1% for

a-Si:H/nc-Si:H tandem cell were reported, very similar efficiencies as compared to the standard

RF PECVD. The whole tandem cell in a batch process is completed within 30 minutes, including

the time consuming doped layers. Once applied to R2R, the tandem cell fabrication with a

reasonable belt size can be accomplished using VHF PECVD. The second development has to be

the width of the foil that needs an upgradation of the R2R machine design. Tests have been made

on 140 cm width foils. Module production with large width foils by Hyet Solar by this technique

is expected in the near future that would increase the power per module. Other lift-off or peel-and-

stick methods for a-Si:H solar cells are glass as temporary substrate [77], colloidal transfer-printing

method [78] and water-assisted transfer printing [76]. The latter showed an efficiency of ~7% for

a-Si:H single junction.

(ii) Low temperature process: There are two areas where adaptation to the conventional

fabrication process of a-Si:H solar cell is done. The first is the textured front (superstrate cell) or

back (substrate cell) surfaces. Normally these are done at high temperatures. In the substrate type

cell, natural texture in Ag is achieved with depositions at >300ºC. Whereas in the superstrate cell,

natural texturing is achieved in TCO (Fluorine doped tin oxide SnO2:F) with deposition at >400ºC.

There are two solutions: (i) Textured substrate - hot-embossing is a very effective technique,

though nanolithography and holographic gratings are also used, especially useful for paper
 33

substrate. The latter two are also explored for high temperature resistant substrates, not necessarily

only for low temperature plastics. The highest efficiency of a-Si:H cell on PEN type substrate was

achieved by nanolithography at EPFL; (ii) Textured TCO – post-deposition texturization is

performed on TCO. For example, Al (also boron, Ga) doped TCO (ZnO) is texture-etched to create

inverse pyramidal crater type surface using diluted (~1%) HCl etching. This type of textured TCO

is also used routinely on glass and other high temperature substrates. Here, the ZnO is deposited

at room temperature by magnetron sputtering (due to particle bombardment, actual substrate

temperature reaches 100ºC during the deposition period). In a typical substrate type cell, 1000 nm

ZnO is deposited by sputtering, followed by texture-etching using HCl, then deposition of ~100

nm Ag, ~100 nm ZnO:Al, followed by the n-i-p a-Si:H cell and top contact (Fig.5). In a superstrate

cell, it is more simple, deposition of 1000 nm of ZnO:Al, followed by texture-etching, then the p-

i-n cell and back contact. An issue with employing thick TCO layer is that it adds high stress

(typically compressive) that forces the foil to curl and it creates cracks in TCO and cell. Symmetric

deposition of thick layers on both sides of substrate, followed by texture etching on one side for

cell fabrication is one of the solutions. Additional precautions to protect the moisture sensitive

ZnO, such as barrier layer between plastic and ZnO (e.g. ZnSnOx) [79], or H/Ar treatment on ZnO

[80] are essential.

 34

Fig.5 Structure of a typical a-Si:H substrate type cell on plastic foil. For PET, PEN, PC type substrates all the
processing is done below 130ºC. Symmetric ZnO:Al layers on both sides of the plastic foil is applied to minimize
curling/stress. The p-type layer is a protocrystalline material made from high hydrogen dilution of silane gas. The n-
type layer is a double layer of phosphous doped a-Si:H/nc-Si:H. Ag is used for grid lines. The intrinsic layer is a-Si:H
made from hydrogen diluted silane. All silicon layers are made by VHF PECVD in a multi-chamber.

(c) Paper substrate: It is a very attractive proposition for low power niche application. Among

papers, cellulose type papers are best suited, as the paper material degrades at temperatures only

above 250ºC. For one type of cellulose paper (FS-2 paper substrates made by Felix Schoeller

Group), it has been shown that planarization by a hydrophilic mesoporous material that mitigates

the macro rough fibrous surface and enhances the wetting property and sticking of back metal

contact is very useful and an efficiency of 3.4% is obtained for a-Si:H cell with n-i-p configuration

[81]. Adding textured back reflector on FS-2 paper through UV nanoimprint lithography, the

efficiency of a-Si:H cell was enhanced to 5.5% [82]. An interesting result is solar cell on plain

white copying paper (80 g/m2) used in office copying machines. The commodity paper was first

planarized by UV curable acrylate lacquer (MO18), followed by another layer of the same lacquer
 35

that was imprinted with a periodic nano-texture. An n-i-p a-Si:H cell made by HWCVD (at 200ºC!)

on this substrate delivered 6.7% efficiency, almost the same efficiency as on SS [83].

(d) Flexible Glass: An extremely thin (~100 μm) glass acts as a flexible foil. The advantage of

this substrate is the excellent encapsulation property with ultra-low WVTR, which is impossible

to achieve with any of the previous two types (metal foils and plastics) of substrates. Moreover,

possibility of superstrate type cell fabrication, easier monolithic integration, optimal temperature

processing, high light transmission, and light weight etc. are the definite plus points. The main

drawback is that these substrates, though bendable, are not rollable and restricted in the extent of

bending. Substrate handling without breakage is also an issue and this poses questions on the up-

scaling of the device. Nevertheless, initial results on small area devices on FTO coated 100 m

thin glass show an encouragement in efficiency (6.95% for a-Si:H and 9.3% for a-Si:H/c-Si:H

cell) [11]. The efficiency of a-Si:H solar cells on various flexible substrates are given in Table 4.

Table 4: Efficiency of some prominent a-Si:H based cells on flexible foil (T transfer method,
*stabilized, # measured under AM0 light condition for space application).

Cell type Substrate Source Eff. (%)

p-i-n (a-Si:H) T Polyester Univ. Utrecht/Nuon, Netherlands 7.7 [84]
p-i-n (a-Si:H) PC 130°C University Utrecht
7.4 [85]
p-i-n (a-Si:H) PS Aarhus University, Denmark
9.5 [86]
n-i-p (a-Si:H) PEN 150°C IMT, (EPFL) Switzerland 8.7 [87]
n-i-p (a-Si:H) E/TD 150°C AIST, Japan 6 [88]
n-i-p (a-Si:H) PET 100°C Univ. Utrecht, Netherlands 5.9 [89]
n-i-p (nc-Si:H) LCP 150°C AIST, Japan 8.1 [90]
p-i-n (a-Si:H/nc-Si:H) T Polyester Helianthos (Hyet Solar) 9.4* [91]
n-i-p (a-Si:H/nc-Si:H) PEN IMT, (EPFL) Switzerland 11.2 (9.8*) [92]
 36

n-i-p (a-Si:H/a-SiGe:H) Kapton® Fuji Electric, Japan 10.1 (9*) [93]

n-i-p (a-Si:H/a- SS Uni-Solar, USA 16.3 [94]
SiGe:H/nc-Si:H)
n-i-p (a-Si:H/a-SiGe:H/a- SS Uni-Solar, USA 13* [95]
SiGe:H)
n-i-p (a-Si:H/a-SiGe:H/a- Polymer Uni-Solar, US 9* [96]
SiGe:H)

5.2 Encapsulation
One of the biggest issues facing plastic substrate based solar cells is the high WVTR and

gas permeation (~1 g/m2/day) , which is not the case with solar cells on glass substrates where the

cell is protected on both sides by a very low WVTR of around 10-12 g/m2/day. The flexible cells

based on SS have this advantage only on the back side (metal), whereas the top side still needs to

be encapsulated with transparent layer with low WVTR. Moreover, they have to protect the plastic

from UV and in addition to being themselves UV resistant. Normally polymer encapsulation, such

as DuPont™ Tedlar® polyvinyl fluoride films as back sheet and DuPont™ Teflon® ETFE or Teflon®

FEP films (fluoropolymer Tefzel) front sheet, along with DuPont “Elvax” EVA resin encapsulant,

is used, which makes the module very expensive. On the other hand, stability of the module on

plastic substrates cannot be promised for a long warranty period (>10 year), making the levelized

cost of electricity (LCOE) price too high. Organic/inorganic multilayer structure can be used to

address this issue.

A WVTR of 5x10-6 g/m2/day has been reported by a three layer (Fig.6) organic/inorganic

structure of a-SiNx/PGMA/a-SiNx (PGMA stands for poly(glycidyl methacrylate)) made by

HWCVD for the inorganic layer and i-CVD (initiated chemical vapor deposition) for the organic

layer [97]. Simulations and modeling can be used to explain the permeation dynamics of such

multilayers and for optimizations of the layer stack [98]. Meyer Burger is supplying an integrated
 37

tool consisting FLEx LT PECVD tool mainly for depositing an inorganic barrier coating and the

PiXDRO JETx inkjet printer for an organic planarization step and claims a WVTR of 10-6

g/m2/day. Single layer Al2O3 by R2R atomspheric atomic layer deposition [99], SiOx by R2R

magnetron sputtering [100], silica-like bilayer by R2R atmospheric pressure-PECVD [101],

SiOxCzHw (organic)/SiOx (inorganic)/SiOxCzHw (organic) layer stack by PECVD [102] are some

of the other options.

Fig.6 Moisture impermeable encapsulation scheme for a-Si:H solar cells on foils. Amorphous SiNx is made by
HWCVD, PGMA is made by i-CVD (a variant of HWCVD). All depositions of the mult-ilayer are done below 110ºC
substrate temperature.

5.3 Degradation
Terrestrial modules: Cells on polymer foil have shown excellent environment stability [103]. The

degradation due to damp heat for 1000 hrs, humidity freeze for 90 cycles, and UV exposure for

1000 hrs is only < 2%, <3% and < 5% respectively. The polymer encapsulated cells of Hyet Solar

and their Powerfoil® products have passed the IEC 61646 and EN 61730 tests (accelerated lifetime

testing, hailstones, cut susceptibility, fire testing) [104], whereas the cells on SS (Unisolar) are
 38

operating in the field for more than 20 years. The light induced degradation (LID) with AM 1.5

light is now <4% for triple junction [105, 106] and double junction [91] and it is comparable to

the LID of such cells on glass substrate [107] and both giving comparable stabilized efficiencies

(~13.6%). However, it should be noted that the stabilized record efficiency (13.6% on SS foil)

[105] is not obtained from the champion initial efficiency (16.3%), which only emphasizes that it

is the absolute value of stabilized efficiency that is important and should be compared with other

type of modules, and not the amount or rate of degradation.

Other applications: The LID with AM0 light for 1000 hrs shows only 6.1% loss for a-Si:H/nc-

Si:H/nc-Si:H and 7.3% for a-Si:H/a-SiGe:H/a-SiGe:H Uni-Solar triple junctions on SS foils [96].

The degradation studies with proton and electron beam show that the radiation tolerance of a-Si:H

solar cells is far superior than that of c-Si solar cells for space use, however, advantages compared

with other types of thin PV cells are still not conclusive. The proton induced degradation is

considerably smaller for a-Si:H and nc-Si:H cells compared with c-Si cells [108] which is

attributed mainly to the thickness of the absorber layer [109].

5.4 Industrial scenario

A-Si:H solar module fabrication is in fact one of the most successful case of flexible solar

cell product. Uni-Solar made high efficiency multi-junctions solar cells on SS foil (6 lines of 2.5

km foil in one run) and ran a very successful business. In the first decade of this millennium, the

R2R production technology became a hot topic [110] and many companies and R&D facilities,

inspired by the success of Uni-Solar, ventured into R2R manufacturing, though for practical

purposes mostly can be categorized as pilot level. PowerFilm and Fuji electric had commercial

production of cells at a smaller scale (<25 MW) using Kapton® as substrates.

AkzoNobel/Nuon/Hyet Solar (Netherlands), Flexcell (VHF electronics, Switzerland) had

 39

interesting products on polyester foils, but could not start commercial production, most probably

due to economic consideration. As far as equipment tools are concerned, there is a very few

possibility to get an off-the-shelf system. The solar cell manufacturing (R&D) companies have

developed their own systems. However, Applied Material (US) has reported R2R PECVD

deposition equipments for solar cell fabrication [111], whereas Meyer Burger has for the

encapsulation, as mentioned earlier.

5.5 Flexible a-Si:H applications and outlook

There are several applications - space application (robustness to radiation damage and light

weight less than 1200 W/Kg) [96], medical application (radiation detector), agriculture (micro-V

textured PC green house), wall paper, food industry (package label), textiles (mobile charger),

outdoor utility (tent), security (features in bank notes, credit cards, passport, art objects), metal-air

batteries etc. The space application is particularly a high investment application which does not

suffer from WVTR problem, however, the radiation damage to the efficiency still is an issue [112].

Other than the niche market, the success will predominantly depends on (i) the cost of

encapsulation compared to cell fabrication and life time warranty, and (ii) the stabilized efficiency

value which is limited to the temperature and stress effect. The latter issue demands that only

multi-junctions will be successful. In that case high deposition rate at low temperature deposition

will remain a challenge. At present the transfer method looks more promising, as the temperature

effect can be avoided. Uni-Solar stopped production and the other flexible solar cell companies

(Fuji, VHF electronics, PowerFilm, Hyet solar etc.) either became inactive or restricted to only

R&D activity. This is due to the traction in business confronting the low-cost c-Si based modules.

However, the flexible a-Si:H cell technology is robust. The Uni-Solar modules are still in operation

after long years (> 20 years) in field. The robustness of the modules is best demonstrated in the
 40

module rolls (Fig. 7) where they remain intact even with friction from rolling over. The amorphous

nature of the material protects its flexibility that is unique to this type of cell. Uninterrupted

kilometer wise solar foils fabrication has been demonstrated by Uni-Solar and Hyet Solar. Hence

throughput and large area applications are certainly an advantage. One particular area where a

successful demonstration of large area light weight a-Si:H modules (Building Applied

Photovoltaics systems, BAPV) is the large IKEA roof tops (Hyet Solar in Netherlands IKEA) on

which c-Si types of solar modules would lead to collapse of roof.

Fig. 7 R2R Helianthos a-Si:H flexible solar cells [113]. Single junction superstrate type of a-S:H cell by PECVD is
deposited by R2R process at standard temperature (as on glass) on 35 cm Al foil roll as temporary substrate and
transferred to a polyester foil roll by a lift-off process.

Among the thin film PV technologies, Si thin film has an advantage when very large (>10

m2) modules are employed, due to the lowest efficiency loss with up scaling [114] and easy

monolithic integration compared to other type of cells. Hyet solar can produce >24 meter long foil

single modules and they claim a 30% reduction in BOS with extreme light weight (600 gr/m²) and
 41

ultra-thin (0.5 mm) modules resulting from almost zero cost of transportation, low cost mechanical

installation (<$0.05/Wp) and electrical installation (<$0.05/Wp) for BIPV foils, and a 35%

reduction in LCOE (a low $0.025/kWh) with their 8% conversion efficiency foils compared to

glass panels (c-Si) with 15% efficiency.

6.0 Comparison between flexible CIGS, CdTe and a-Si:H cells and applications
In terms of stoichiomety, producing CdTe and a-Si:H solar cells is easier than CIGS solar

cells. Stability of CIGS and CdTe based cells are higher than that of a-Si:H cells. For high

efficiency, CIGS and CdTe are suitable candidates however at present on flexible foils CIGS cells

have shown high efficiency than CdTe cells, due to the fact that CIGS is built-in substrate

configuration, which is the best structure for flexible photovoltaics. Up scaling has been already

demonstrated for CIGS cells, with some limitation on the type of substrates to be used. CdTe solar

cells have the simplest stoichiometry that helps to improve industrial scalability; on the other hand

the best configuration for this technology is superstrate configuration (about 22% vs 13.6% in

substrate configuration). This needs the application of a flexible substrate that has to be also

extremely transparent, leaving very small choice for a suitable substrate (the best is the ultra thin

glass solution). The a-Si:H type is good at low cost, easily scalable, long single module can be

used to wrap across a complete big roof. The biggest advantage is that a large number of substrates

can be used. Hence there are unlimited types of applications as mentioned below.

Flexible modules can be fixed by glue on the metal roof and automobile roof and mounting

racks are not required for installation A 60W flexible module is used to power camping equipments

such as headlamps, tent etc. Flexible modules are also used in GPS systems, recharging field

communication radios, mobile phone, laptop computers, sensors and night vision goggles. Flexible

cells are integrated into dress and backpacks to generate power (~1W) that can be used to energize
 42

portable devices, for example mobile phone. Flexible panel installed over the vehicle roof is used

to charge batteries.

7.0 Summary and Conclusions

Flexible CIGS, CdTe and a-Si:H thin film solar cells are reviewed. Thin film solar cells

have acquired a competitive photovoltaic market in the past years against the formidable Si

photovoltaics which has a dominant market share of around 90%. The reason for the sustaining

existence of thin film cells even in a non-favorable eco system is due to its specialty such as less

material wastage and low thermal budget processing etc. Thin films are an ideal type to fabricate

on flexible substrates due to their ability to adapt to the shape of the substrate, to their robustness

and their high absorption coefficient. The potential of flexible thin film PV is beyond the limited

market, it has the potential, with well-established technology to large production volumes,

combining high throughput and large area devices.

For polymer foils, absorption in the low wavelength region is not negligible. One of the

biggest issues with polymer foils is the high WVTR and gas permeation of about 1 g/m2/day. This

is not the case with glass substrates where the cell is protected on both sides by a very low WVTR

of ~10-12 g/m2/day. The ultra-thin (~100μm) glass solves these issues; its transparency is the same

as the rigid glass. It has an excellent encapsulation property with ultra-low WVTR which is not

possible to achieve with metal and plastic foils. However, the cost of ultra-thin flexible glass is

high. Though flexible glass substrates are bendable, they are not rollable and they are restricted in

the extent of bending.

In terms of efficiency and stability, CIGS solar cells are comparable to Si solar cells.

However, the main issues are lower yield, lower manufacturing capacity and higher module

production cost. High temperature processing of CIGS solar cells is an issue for plastic foils,
 43

whereas for metal foils, impurity diffusion from foils to the device is an issue. To avoid this,

diffusion barriers such as Al2O3, Si3N4 etc., are used in flexible CIGS cells. High efficiency CIGS

devices are fabricated by co-evaporation to tailor the film composition for controlled bandgap

grading. Using three stage deposition followed by post-deposition (NaF and KF) treatment at

<400˚C, highest efficiency of 20.4% has been reported on polyimide. One of the main concerns is

the protection from environment and flexible encapsulation with life-span of 20 years is required.

CdTe solar cell performs better in superstrate configuration. For polyimide foils in

superstrate configuration, its stability against UV radiation is an issue; an anti-UV coating on the

other side of the substrate is required. For monolithic integration, plastic foils are preferred over

metallic foils due to an extra insulating layer for metal foils. Flexible glass and flexible ceramics

do not have such issues; however, the cost is the problem. For substrate devices, tiles and ceramics

can be used. For both superstrate and substrate type cells, stability during bending is required.

Simple stoichiometry and high reproducibility of CdTe solar cells reduce the problems with

inhomogeneity which are more frequent in CIGS solar cells.

For a-Si:H solar cells, apart from standard foils and flexible glass, other notable substrates

are papers such as cellulose papers, bank notes and security papers. Paper substrate is very

attractive for low power application. Among papers, cellulose papers are suitable since the material

degrades >250ºC. Planarization by a hydrophilic material reduces the rough surface of the paper

and improves the wetting property useful for back contact metal. About 6.7% efficiency is obtained

for single junction on plain white copying paper which is almost equal to the efficiency on SS

foils. High efficiency multi-junction solar cells on a long SS foil has been reported by Uni-Solar

(6 lines of 2.5 km foil in one run). In standard PECVD process (RF 13.56 MHz), 36 cm width foil

is used by the industry in pilot module production and further development is required. Since the
 44

belt size in R2R fabrication depends on the deposition rate, a high deposition rate is required

particularly for the tandem cell that requires a thick (1500 nm) nc-Si:H bottom cell. VHF PECVD

was explored for high deposition rate and similar efficiencies have been reported as compared to

the standard PECVD. The success of flexible a-Si:H cells strongly depends on the cost of

encapsulation and stabilized efficiency. For encapsulation, a three layer hybrid structure (a-

SiNx/PGMA/a-SiNx) can be used due to its very low WVTR of 5x10-6 g/m2/day. Since polymer

foils are moisture sensitive, Si containing moisture barriers such as SiOx, SiNx and SiOxNy are

used. Plasma treatment improves the adhesion between polymer foil and moisture barrier. Flexible

a-Si:H cell technology is robust. Uni-Solar modules are still in operation for more than 20 years in

the field. The robustness is best demonstrated in the module rolls even with friction from rolling

over. The amorphous nature of the material protects its flexibility that is unique to this type of cell.

An emerging solar cell technology is flexible perovskite solar cells based on hybrid

organic-inorganic perovskite thin film and it offers low cost power production. Perovskite absorber

materials can be deposited by inexpensive techniques such as spin-coating, spray-coating etc. Its

suitable band gap of 1.52 eV, high absorption co-efficient (105/cm), bending stability and

maintaining efficiency even after bending tests makes it suitable for roll-to-roll production line.

The main issue of perovskite solar cell is moisture absorption leading to degradation of perovskite

absorber layer and poor long-term stability. If the moisture induced degradation issues are solved,

apart from flexible CIGS, CdTe and a-Si:H modules, flexible perovskite modules can also be used

successfully in BIPV, roof-top, automobile and wearable electronics (for example health monitor)

applications.

In conclusion, as compared to conventional Si solar cells, efficiency of flexible CIGS cell

is comparable and stability is not a concern. Less energy input is required to deposit CIGS film
 45

than that of Si. The other technology flexible CdTe cell has demonstrated successful industrial

production due to simple stoichimetry, high reproducibility, high efficiency and fabrication in both

superstrate and substrate configuration. In the case of flexible a-Si:H, since PECVD is the standard

technique used for fabrication with processing temperature around 200°C, temperature sensitive

plastic foils can be used. The success of flexible a-Si:H cell is due to high throughput, large area,

and long time PECVD deposition without down time. Therefore, future progress is bright for

flexible CIGS, CdTe and a-Si:H solar cells. Bridging the efficiency gap between rigid and flexible

cells is feasible and the gap can be narrowed by applying the knowledge pioneered on glass.

References
[1] Tatsuo, S. Advances in crystalline silicon solar cell technology for industrial mass

production. NPG Asia Materials 2, 96–102 (2010).

[2] Tiwari, A. N., Romeo, A., Baetzner, D. & Zogg, H. Flexible CdTe solar cells on polymer

films, Prog. Photovolt: Res. Appl. 9, 211-215 (2001).

[3] Brémaud, D. J. L. Investigation and Development of CIGS Solar Cells on Flexible

Substrates and with Alternative Electrical Back Contacts. Ph.D. thesis, 2009, p.14.

[4] Dou, B. et al. High-Performance Flexible Perovskite Solar Cells on Ultrathin Glass:

Implications of the TCO. J. Phys. Chem. Lett. 8, 4960−4966 (2017).

[5] Reinhard, P. et al. High efficiency flexible Cu(In,Ga)Se2 solar cells, 20th International

Workshop on Active-Matrix Flat panel Displays and Devices. 20, 79-81 (2013).

[6] Jeyakumar R. & Singh, U. P. Copper indium gallium selenide based solar cells - a review.

Energy Environ. Sci. 10, 1306-1319 (2017).

 46

[7] Romain, C. et al. Advanced Alkali Treatments for High‐Efficiency Cu(In,Ga)Se2 Solar
Cells on Flexible Substrates, Advanced Energy Materials 9, 1900408 (2019)

[8] EMPA Thin Films and Photovoltaics, March 2015.

[9] Mahabaduge, H. P. et al. High-efficiency, flexible CdTe solar cells on ultra-thin glass

substrates. App. Phys. Lett. 106, 133501-4 (2015).

[10] Jeyakumar R. & Verma, A. Photovoltaic Properties of a-Si:H Films Grown by Plasma

Enhanced Chemical Vapor Deposition: A Review. Mater. Express 2, 177-196 (2012).

[11] Myong, S. Y., Jeon, L. S. & Kwon, S. W. Superstrate type flexible thin-film Si solar cells

using flexible glass substrates. Thin Solid Films 550, 705–709 (2014).

[12] Chirila, A. et al. Highly efficient Cu(In,Ga)Se2 solar cells grown on flexible polymer films.

Nat. Mater. 10, 857-861 (2011 ).

[13] Feurer, T. et al. Progress in thin film CIGS photovoltaics - Research and development,

manufacturing, and applications. Prog. Photovoltaics Res. Appl. 25, 645–667 (2017).

[14] Todorov, T. K., Gunawan, O., Gokmen, T. & Mitzi, D. B. Solution-processed

Cu(In,Ga)(S,Se)2 absorber yielding a 15.2% efficient solar cell. Prog. Photovoltaics Res.

Appl. 21, 82-87 (2012).

[15] Hibberd, C. J. et al. Non-vacuum methods for formation of Cu(In, Ga)(Se, S)2 thin film

photovoltaic absorbers. Prog. Photovoltaics Res. Appl., 18, 434–452, 2009.

[16] Rui, K. New World Record Cu(In,Ga)(Se,S)2 Thin Film Solar Cell Efficiency Beyond 22%.

Proc. IEEE 43rd Photovolt. Spec. Conf. 43, 3-7 (2016).

[17] Reinhard, P. et al. Features of KF and NaF Postdeposition Treatments of Cu(In,Ga)Se2

Absorbers for High Efficiency Thin Film Solar Cells. Chem. Mater. 27, 5755–5764 (2015).

[18] Forest, R. V. et al. Reversibility of (Ag,Cu)(In,Ga)Se2 electrical properties with the addition
 47

and removal of Na: Role of grain boundaries. J. Appl. Phys. 117, 115102 (2015).

[19] Jyh-Lih Wu, Yoshiaki Hirai, Takuya Kato, Hiroki Sugimoto, Veronica Bermudez, New

World Record Efficiency up to 22.9% for Cu(In,Ga)(Se,S)2 Thin-Film Solar Cells, 7th

World Conference on Photovoltaic Energy Conversion, June 10-15, 2018 Waikoloa,

Hawaii, USA.

[20] T. Kato, Cu(In,Ga)(Se,S)2 solar cell research in solar frontier: progress and current status

temperatures. Jpn. J. Appl. Phys. 56, 04CA02 (2017).

[21] Jackson, P. et al. Effects of heavy alkali elements in Cu(In,Ga)Se2 solar cells with

efficiencies up to 22.6%. Phys. Status Solidi - Rapid Res. Lett. 4, 1-4 (2016).

[22] Jackson, P. et al. Compositional investigation of potassium doped Cu(In,Ga)Se2 solar cells

with efficiencies up to 20.8%. Phys. Status Solidi - Rapid Res. Lett. 8, 219–222 (2014).

[23] Chirilă, A. et al. Potassium-induced surface modification of Cu(In,Ga)Se2 thin films for

high-efficiency solar cells. Nat. Mater. 12, 1107-1111 (2013).

[24] Farshchi, R., Hickey, B. & Poplavskyy, D. Light-Soak and Dark-Heat Induced Changes in

Cu(In,Ga)Se2 Solar Cells: A Macroscopic to Microscopic Study. Proc. 44th IEEE PVSC,

2017, pp. 1-4.

[25] Kaczynski, R. et al. In-line Potassium Fluoride Treatment of CIGS Absorbers Deposited on

Flexible Substrates in a Production-Scale Process Tool. Proc. 44th IEEE PVSC, 2017, pp-

1-4.

[26] Reinhard, P. et al. Review of progress toward 20% efficiency flexible CIGS solar cells and

manufacturing issues of solar modules. IEEE J. Photovoltaics 3, 572–580 (2013).

[27] Ishizuka, S. et al. Alkali incorporation control in Cu(In,Ga)Se2 thin films using silicate thin

layers and applications in enhancing flexible solar cell efficiency. Appl. Phys. Lett. 93,
 48

124105 (2008).

[28] Kenichi, M. et al. Monolithically integrated flexible Cu(In,Ga)Se2 solar cells and
submodules using newly developed structure metal foil substrate with a dielectric layer,
Solar Energy Materials & Solar Cells 112, 106–111 (2013).

[29] Kessler, F. & Rudmann, D. Technological aspects of flexible CIGS solar cells and modules.

Sol. Energy, 77, 685–695 (2004).

[30] Urbaniak, A. et al. Effects of Na incorporation on electrical properties of Cu(In,Ga)Se2-

based photovoltaic devices on polyimide substrates. Sol. Energy Mater. Sol. Cells, 128, 52–

56 (2014).

[31] Kim, K., Eser, E. & Shafarman, W. N. Na Incorporation in Cu(In,Ga)(Se,S)2 films grown

on insulator-coated stainless steel foil using a metal precursor reaction. IEEE J.

Photovoltaics 5, 1222–1228 (2015).

[32] Eisenbarth, T. et al. Influence of iron on defect concentrations and device performance for

Cu(In,Ga)Se2 solar cells on stainless steel substrates. Prog. Photovoltaics Res. Appl. 20,

568–574 (2012).

[33] Guo, L. et al. A Diffusion Barrier for Flexible Thin Film Photovoltaics. J. Electrochem. Soc.

160, D102–D106 (2013).

[34] Batchelor, W. K. et al. Substrate and back contact effects in CIGS devices on steel foil,
Proc. 29th IEEE PVSEC, 29, 716–719 (2002).

[35] Pianezzi, F. et al. Electronic properties of Cu(In,Ga)Se2 solar cells on stainless steel foils

without diffusion barrier. Prog. Photovoltaics Res. Appl. 20, 253–259 (2012).

[36] Abou-Ras, D. et al. Innovation highway: Breakthrough milestones and key developments

in chalcopyrite photovoltaics from a retrospective viewpoint. Thin Solid Films, 633, 2-12,

2017.
 49

[37] Margolis, R., Feldman, D. & Boff, D. Q1/Q2 2017 Solar Industry Update 2017.

[38] Lee, J. et al. Effect of Three-Stage Growth Modification on a CIGS Microstructure. IEEE

J. Photovoltaics 6, 1645–1649 (2016).

[39] Shogo, I. et al. Monolithically integrated flexible Cu(In,Ga)Se2 solar cell submodules,
Solar Energy Materials & Solar Cells 94, 2052–2056 (2010).

[40] Arnaud, G. et al. CIGS solar cells on flexible ultra-thin glass substrates: Characterization
and bending test, Thin Solid Films 592, 99–104 (2015)

[41] Karsten, O., Liudmila, M., Alexander, B., Igor, K., Flexible Cu(In,Ga)Se2 thin-film solar
cells for space application, Thin Solid Films 511-512, 613-622 (2006).

[42] Hansung, K., Da, X., Ciby, J., and Yaqiong, W., Modeling Thermo-Mechanical Stress of
Flexible CIGS Solar Cells, IEEE J. Photovoltaics 9, 490-505 (2019).

[43] Yi, C. L., Residual stress in CIGS thin film solar cells on polyimide: simulation and
experiments, Journal of Materials Science: Materials in Electronics, 25 (2014) 461–465.

[44] Paudel, N. R., Wieland, K. A. & Compaan, A. D. Ultrathin CdS/CdTe solar cells by

sputtering. Sol. Energy Mater. Sol. Cells 105, 109-112 (2012).

[45] Paudel, N. R., Compaan, A. D. & Yan, Y. Ultrathin CdTe solar cells with MoO3-x/Au back

contacts. J. Electron. Mater. 43, 2783-2787 (2014).

[46] Plotnikov, V. V., Kwon, D., Wieland, K. A. & Compaan, A. D. 10% efficiency solar cells

with 0.5 µm of CdTe. 34th IEEE Photovolt. Spec. Conf. 34, 1435-1438 (2009).

doi:10.1109/PVSC.2009.5411312.

[47] Moutinho, H.R., Al-Jassim, M. M., Abufoltuh, F.A., Levi, D.H., Dippo, P.C., Dhere,

R.G., Kazmerski, L.L., Studies of Recrystallization of CdTe Thin Films After CdCl 2

Treatment, 26th IEEE Photovoltaic Specialists Conference, 29 Sept. – 3 Oct., 1997,

Anaheim, USA, pp. 431-434.

 50

[48] Xu, B.L., Rimmaudo, I., Salavei, A., Piccinelli, F., Di Mare, S., Menossi, D., Bosio, A.,

Romeo, N., Romeo, A., CdCl2 activation treatment: A comprehensive study by monitoring

the annealing temperature. Thin Solid Films 582, 110–114 (2015).

[49] Halliday, D., Eggleston, J., Durose, K., A photoluminescence study of polycrystalline

thin-film CdTe/CdS solar cells. J. Cryst. Growth 186, 543–549 (1998).

[50] Perrenoud, J., Kranz, L., Gretener, C., Pianezzi, F., Nishiwaki, S., Buecheler, S., Tiwari,

A.N., A comprehensive picture of Cu doping in CdTe solar cells, J. Appl. Phys. 114,

174505 (2013).

[51] Laurenti, J.P., Bastide, G., Rouzeyre, M., Localized Defects in p-CdTe:Cu doped by

copper incorporation during bridgman growth. Solid State Commun. 67, 1127–1130

(1988).

[52] Chou, H.C., Rohatgi, A., Jokerst, N.M., Thomas, E.W., Kamra, S., Copper migration in

CdTe heterojunction solar cells, J. Electron. Mater. 25, 1093–1098 (1996).

[53] Kranz, L. et al. Doping of polycrystalline CdTe for high-efficiency solar cells on flexible

metal foil. Nat. Commun. 4, 1-7 (2013).

[54] Williams, B. L. et al. Challenges and prospects for developing CdS/CdTe substrate solar

cells on Mo foils. Sol. Energy Mater. Sol. Cells 124, 31-38 (2014).

[55] Bittau, F., Artegiani, E., Abbas, A., Menossi, D., Romeo, A., Bowers, J.W., Walls, J. M.,

Magnesium-doped Zinc Oxide as a High Resistance Transparent Layer for thin film CdS /

CdTe solar cells, 44th IEEE Photovolt. Spec. Conf., Washington, D.C., June 25-30, 2017,

pp. 752-756.

[56] Sites, J., Munshi, A., Kephart, J., Swanson, D., Sampath, W.S., Progress and Challenges

with CdTe Cell Efficiency, 43rd IEEE Photovolt. Spec. Conf., Portland, OR, USA, June
 51

5-10, 2016, pp. 3632-3635.

[57] Kephart, J.M., McCamy, J.W., Ma, Z., Ganjoo, A., Alamgir, F.M., Sampath, W.S., Band

alignment of front contact layers for high-efficiency CdTe solar cells, Sol. Energy Mater.

Sol. Cells 157, 266–275 (2016).

[58 ] Munshi, A.H., Kephart, J.M., Abbas, A., Shimpi, T.M., Barth, K.L., Walls, J.M.,

Sampath, W.S., Polycrystalline CdTe photovoltaics with efficiency over 18% through

improved absorber passivation and current collection, Sol. Energy Mater. Sol. Cells 176,

9–18 (2018).

[59] Dhere, R. G. et al. Development of substrate structure CdTe photovoltaic devices with

performance exceeding 10%. 38th IEEE Photovoltaic Specialists Conference 38, 3208-

3211 (2012). doi:10.1109/PVSC.2012.6318260.

[60] Gretener, C. et al. Development of MoOx thin films as back contact buffer for CdTe solar

cells in substrate configuration. Thin Solid Films 535, 193-197 (2013).

[61] Romeo, A. et al. High-efficiency flexible CdTe solar cells on polymer substrates. Sol.

Energy Mater. Sol. Cells 90, 3407-3415 (2006).

[62] Perrenoud, J., Schaffner, B., Buecheler, S. & Tiwari, A. N. Fabrication of flexible CdTe

solar modules with monolithic cell interconnection. Sol. Energy Mater. Sol. Cells 95, S8-

S12 (2011).

[63] Perrenoud, J. C. Low temperature grown CdTe thin film solar cells for the application on

flexible substrates. Ph.D. Thesis (2012). doi:10.3929/ethz-a-007339371.

[64] Rance, W. L. et al. 14%-efficient flexible CdTe solar cells on ultra-thin glass substrates.

Appl. Phys. Lett. 104, 143903-4 (2014).

 52

[65] Gessert, T.A., Romero, M.J., Dhere, R.G., and Asher, S.E., Analysis of the ZnTe:Cu
Contact on CdS/CdTe Solar Cells. 2003 MRS Proc., vol. 763, DOI: 10.1557/PROC-763-
B3.4.

[66] Romeo, N., Bosio, A., Mazzamuto, S., Romeo, A., and Vaillant Roca, L., High Efficiency

CdTe/CdS Thin Film Solar Cells with a Novel Back Contact, Proc. 22nd EU PVSEC, 2007,

1919–1921.

[67] Artegiani, E, Menossi, D, Shiel, H, et al. Analysis of a novel CuCl2 back contact process

for improved stability in CdTe solar cells. Prog Photovolt Res Appl. 27, 706– 715 (2019).

[68] Salavei, A. et al. Comparison of high efficiency flexible CdTe solar cells on different

substrates at low temperature deposition. Sol. Energy 139, 13-18 (2016).

[69] http://annals.fih.upt.ro

[70] Wipliez, L. et al. Monolithic Series Interconnection of Flexible Thin-Film PV Devices. 26th

European Photovoltaic Solar Energy Conference and Exhibition, 5-9 September 2011,

Hamburg, Germany. pp. 2641-2644.

[71] Soppe, W. et al. Phys. Status Solidi A 210, 707-710 (2013).

[72] Tsao, Y.-C. et al. Appl. Phys. A 120, 417–425 (2015).

[73] Xiao, H. et al. Performance optimization of flexible a-Si:H solar cells with nanotextured

plasmonic substrate by tuning the thickness of oxide spacer layer. Nano Energy 11, 78-87

(2015).

[74] Dong, W. J. et al. Monolithic Photoassisted Water Splitting Device Using Anodized Ni-Fe

Oxygen Evolution Catalytic Substrate Adv. Energy Mater. 7, 1700659-7 (2017).

[75] Lin, Y. et al. Dual-Layer Nanostructured Flexible Thin-Film Amorphous Silicon Solar

Cells with Enhanced Light Harvesting and Photoelectric Conversion Efficiency ACS

Appl. Mater. Interfaces 8, 10929-10936 (2016).

 53

[76] Lee, C.H., Kim, D. R. & Zheng, X. Transfer Printing Methods for Flexible Thin Film

Solar Cells: Basic Concepts and Working Principles ACS Nano 8, 8746–8756 (2014).

[77] Wang, G. H. et al. Fabrication of flexible silicon thin film solar modules by substrate

transfer technology. Vacuum 102, 72-74 (2014).

[78] Yang, X. et al. Colloidal transfer printing method for periodically textured thin films in

flexible media with greatly enhanced solar energy harvesting. Mater. Res. Express 2

106402 (2015).

[79] Wilken, K., Finger, F. & Smirnov V. Phys. Influence of ZnSnOx barrier layer on the

texturing of ZnO:Al layers for light management in flexible thin-film silicon solar cells.

Status Solidi A 214, 1600884-7 (2017).

[80] Jae-Seong Jeong, J. -S. et al. Effect of H/Ar treatment on ZnO: B transparent conducting

oxide for flexible a-Si:H/μc-Si:H photovoltaic modules under damp heat stress.

Microelectronics Reliability 64, 640–645 (2016).

[81] Águas, H. et al. Thin Film Silicon Photovoltaic Cells on Paper for Flexible Indoor

Applications. Adv. Funct. Mater. 25, 3592–3598 (2015).

[82] Smeets, M. et al. Flexible thin film solar cells on cellulose substrates with improved light

management. Phys. Status Solidi A 214, 1700070-5 (2017)

[83] van der Werf, C. H. M. et al. Amorphous silicon solar cells on nano-imprinted

commodity paper without sacrificing efficiency. Phys. Status Solidi RRL 9, 622–626

(2015)

[84] Rath, J. K. et al. Thin film silicon modules on plastic superstrates. Journal of Non-

Crystalline Solids 354, 2381-2385 (2008).

 54

[85] de Jong, M. M., Rath, J. K. & Schropp, R. E. I. Very Thin Micromorph Tandem Solar

Cells Deposited at Low Substrate Temperature. Mater. Res. Soc. Symp. Proc. 1426, 45-

49 (2012).

[86] Xu, X. et al. High efficiency ultra lightweight a-Si:H/a-SiGe:H/a-SiGe:H triple-junction

solar cells on polymer substrate using roll-to-roll technology. Proc. 33rd IEEE

Photovoltaic Specialists Conference 33, 1-6 (2008).

[87] Söderström, T., Haug, F. –J., Terrazzoni-Daudrix, V. & Ballif, C. Optimization of

amorphous silicon thin film solar cells for flexible photovoltaics. J. App. Phys. 103,

114509-8 (2008).

[88] Mase, H., Kondo, M. & Matsuda, A. Microcrystalline silicon solar cells fabricated on

polymer substrate. Solar Energy Materials and Solar Cells 74, 547-552 (2002).

[89] Rath, J. K. et al. Fabrication of thin film silicon solar cells on plastic substrate by very

high frequency PECVD. Solar Energy Materials and Solar Cells 94, 1534–1541 (2010).

[90] Takeda, T., Kondo, M. & Matsuda, A. Thin film solar cells on liquid crystal polymer.

Pro. 3rd World Conference on Photovoltaic Energy Conversion 3, 1580-1583 (2003).

[91] Hamers, E. A. G. et al. Transfer of 9% a-Si:H/μc-Si:H Tandem Technology to Roll-to-

Roll Production for Large Area Flexible PV Laminates. Proc. 26th EU-PVSEC, Hamburg,

Germany 26, 2379–2383 (2011).

[92] Meillaud, F. et al. Realization of high efficiency micromorph tandem silicon solar cells

on glass and plastic substrates: Issues and potential. Solar Energy Materials and Solar

Cells 95, 127–130 (2011).

[93] Tsuji, T. et al. Large-Area, Light-Weight Flexible Solar Cells Using Plastic Film

Substrates. Proc. 25th EU-PVSEC, 25, 2740-2743 (2010).

 55

[94] Yan, B. et al. Innovative dual function nc-SiOx:H layer leading to a >16% efficient

multi-junction thin-film silicon solar cell. Applied Physics Letters 99, 113512-3 (2011).

[95] Yang, J., Banerjee, A. & Guha, S. Triple-junction amorphous silicon alloy solar cell with

14.6% initial and 13.0% stable conversion efficiencies. Appl. Phys. Lett. 70, 2975 (1997).

[96] Banerjee, A. et al. Advances in cell efficiency of a-Si:H and nc-Si:H-based multi-

junction solar cells for space and near-space applications. Proc. 35th IEEE Photovoltaic

Specialist Conference 35, 2651–2655 (2010).

[97] Spee, D., van der Werf, K., Rath, J. K & Schropp, R. Excellent organic/inorganic

transparent thin film moisture barrier entirely made by hot wire CVD at 100°C. Phys.

Status Solidi RRL 6, 151-153 (2012).

[98] Toni, E., Effects of random defect distributions in the barrier coating on the gas

permeability of multilayer films. Surface & Coatings Technology 302, 65-74 (2016).

[99] Ali, K., Choi, K. -H., Jo, J. & Lee, Y. W. High rate roll-to-roll atmospheric atomic layer

deposition of Al2O3 thin films towards gas diffusion barriers on polymers. Materials

Letters 136, 90–94 (2014).

[100] Bang, S. -H., Hwang, N. -M. & Kim, H. -L. Permeation barrier properties of silicon oxide

films deposited on polyethylene terephthalate (PET) substrate using roll-to-roll reactive

magnetron sputtering. Microelectronic Engineering 166, 39-44 (2016).

[101] Elam, F. M. et al. Atmospheric pressure roll-to-roll plasma enhanced CVD of high

quality silica-like bilayer encapsulation films. Plasma Processes and Polymers 14,

1600143 (2017).
 56

[102] Charifou, R. et al. SiOx and SiOxCzHw mono- and multi-layer deposits for improved

polymer oxygen and water vapor barrier properties. Journal of Membrane Science 500,

245–254 (2016).

[103] Schlatmann, R. et al. Roll-to-roll production of thin film silicon PV modules. Proc. 21st

European Photovoltaic Solar Energy Conference, Dresden, Germany 1587-1589 (2006).

[104] Jäger, K., Lenssen, J., Veltman, K. & Hamers, E. Large-area production of highly

efficient flexible light-weight thin-film silicon PV modules, Proc. 28th European

Photovoltaic Solar Energy Conference and Exhibition, Paris, France 2164-2169 (2013).

[105] Guha, S., Yang, J. & Yan, B. High efficiency multi-junction thin film silicon cells

incorporating nanocrystalline silicon. Solar Energy Materials & Solar Cells 119, 1-11

(2013).

[106] Schropp, R. E. I. et al. Triple Junction n-i-p Solar Cells with Hot-Wire Deposited

Protocrystalline and Microcrystalline Silicon. Mater. Res. Soc. Symp. Proc. 0989-A15-

03, 989 (2007).

[107] Sai, H. et al. Triple-junction thin-film silicon solar cell fabricated on periodically textured

substrate with a stabilized efficiency of 13.6%. Appl. Phys. Lett. 106, 213902-4 (2015).

[108] Sato, S.-I. et al. Temperature influence on performance degradation of hydrogenated

amorphous silicon solar cells irradiated with protons. Prog. Photovolt: Res. Appl. 21,

1499–1506 (2013).

[109] Verkerk, A.D., Rath, J. K. & Schropp, R. E. I. Degradation of thin film nanocrystalline

silicon solar cells with 1 MeV protons. Energy Procedia 2, 221–226 (2010).

[110] Jeyakumar, R. et al. Inorganic photovoltaics-Planar and nanostructured devices. Prog. Mat.

Sci. 82, 294-404 (2016).

 57

[111] Morrison, N. A. et al. An Overview of Process and Product Requirements for Next

Generation Thin Film Electronics, Advanced Touch Panel Devices, and Ultra High

Barriers. Proceedings of the IEEE 103, 518-534 (2015).

[112] Sato, S.-I., Beernink, K. & Ohshima, T. Charged particle radiation effects on flexible a-

Si/a-SiGe/a-SiGe triple junction solar cells for space use. Proc. 39th IEEE Photovoltaic

Specialists Conference, PART 2, 39, 76-82 (2013).

[113] www.nuondoet.nl/#/boog/helianthos/detail

[114] Hamers, E. & Smets, A.

http://cdn.instantmagazine.com/upload/12499/10hamer.35632bd32e12.pdf
 58

Declaration of interests

The authors declare that they have no known competing financial interests or personal

relationships that could have appeared to influence the work reported in this paper.